While biometric genetic studies have long indicated substantial heritability for internalizing disorders, including major depressive and generalized anxiety disorders, progress in mapping this influence onto specific genetic variants has been slow. Expert consensus suggests that finding this missing heritability will require moving beyond the current one SNP at a time genomewide association study (GWAS) paradigm to develop interactive models jointly considering effects across structural and functional genomic units (e.g., genes and biological pathways), as well as modeling gene-environmental interaction (GxE) in the highly dimensional GWAS context. The overarching goal of this proposal is to provide the candidate the training and research opportunities necessary to advance such an interactive model of internalizing disorder etiology, thus supporting the candidate's long-term career goal of becoming an independent investigator in the area. General training goals include building proficiency in statistical genetics and bioinformatics, and strengthening knowledge of substantive issues in developmental psychopathology and psychometric statistical modeling/programming. The candidate proposes to acquire a number of specific skills, including data integration techniques for including prior information in GWAS, aggregation methods to capture the possible effects of many markers with very small effects (e.g., gene-based, pathway-based, and polygenic risk profile score analyses) and machine learning applications for investigating epistasis and nonlinear genomic effects. Proposed training in these areas consists of an interlocking program of coursework, intensive mentoring, summer programs, reading groups, seminar series and conferences, with special attention to training in research ethics. Direct mentoring is a key feature of this program, with access to leading experts in each of the proposed training areas (i.e., Drs. Edwin van den Oord and Patrick Sullivan - statistical genetics and bioinformatics, Dr. E. Jane Costello - developmental psychopathology, and Dr. Michael Neale - statistical modeling and programming) representing a core strength of the proposed training plan. The candidate proposes to apply acquired skills in the research portion of the project by conducting a multistage GWAS to investigate genomic influence, both direct and in interaction with environmental risk factors, on developmental trajectories of internalizing disorders. Methodologically, this approach focuses on integrating longitudinal statistical models of phenotypic development and environmental risk with emerging genomic methods. This will be facilitated by bringing together an unprecedented combination of longitudinal GWAS datasets, including a meta-analysis of the four Duke/VCU samples of the Gene, Environment, and Development Initiative (GEDI) (N=3,623). This discovery phase meta-analysis will then be followed by replication in two large independent longitudinal samples - the National Longitudinal Study of Adolescent Health (N=~12,000) and the Minnesota Twin and Family Study (N=3,762). The exceptional statistical power of this design will be further complimented in a final analytical stage, with next-generation sequencing follow-up using targeted capture methods to extract promising genes/regions and extreme-trait methods to maximize power by selecting subjects from the extremes of the phenotypic distributions. The institutional environment provided by the Virginia Commonwealth University Center for Biomarker Research and Personalized Medicine (CBRPM) represents another core asset to the candidate's career development. The CBRPM's mission is to develop and apply novel methods for the purpose of identifying and using biomarkers to improve disease understanding and medical treatment. The Center operates from a strong multidisciplinary perspective with primary applications involving psychiatric outcomes, developmental psychopathology, and substance use disorders. Currently funded research programs at the CBRPM include serving as the data analysis core of the Duke/VCU GEDI project, method development in the design and analysis of adaptive multistage GWAS, a replication study of findings from schizophrenia GWAS, and whole genome profiling to detect schizophrenia methylation markers. These projects offer an ideal context for executing the proposed training and research, as they provide an environment focused on GWAS and next-generation sequencing applications to psychiatric disorders - central topics of the candidate's career development plan. The CBRPM also provides several unique physical resources supporting the current proposal, including a SOLiD 4 next-generation sequencer, which will allow follow-up sequencing to be done in-house, and a Center-dedicated computing cluster, facilitating the computationally intensive analyses proposed. In addition to physical resources, the CBRPM offers a challenging and collegial intellectual environment that promotes collaboration both within and outside of the Center. In sum, the CBRPM provides a stimulating and supportive institutional environment with the all of the statistical, computational and laboratory resources necessary to make this proposal a successful one.